Dominant parameter selection in the marginally identifiable case

Often a rather limited set of experimental data is available for the identification of a dynamic model, which contains many parameters. This is, e.g. the usual case for crop growth models. In this situation, only some parameter values can be estimated. Based on an analysis of the Fisher information matrix, a method for a reasonable selection of parameters is suggested here. The method chooses the most sensitive parameters, i.e. those to which the model under the considered experimental conditions is most sensitive, and excludes both coupled parameters and those that exhibit multiplecorrelation. A comparison with different ridge regression methods is made. The methodology is illustrated with a simple lettuce growth model.

[1]  F. Buwalda,et al.  Nitrate concentration in greenhouse lettuce: a modelling study , 1998 .

[2]  Raphael Linker,et al.  Modeling the effect of abrupt changes in nitrogen availability on lettuce growth, root–shoot partitioning and nitrate concentration , 2005 .

[3]  Jan Van Impe,et al.  Solving Dual Optimization Problems in Identification and Performance of Fed-Batch Bioreactors , 2000 .

[4]  Ilya Ioslovich,et al.  SE—Structures and Environment: Fitting the Nicolet Lettuce Growth Model to Plant-spacing Experimental Data , 2002 .

[5]  François Gastal,et al.  MecaNiCAL, a supply–demand model of carbon and nitrogen partitioning applied to defoliated grass: 2. Parameter estimation and model evaluation , 1998 .

[6]  Cleve Moler,et al.  Mathematical Handbook for Scientists and Engineers , 1961 .

[7]  Min Zou,et al.  Sensitivity analysis for the design and operation of a non-contacting mechanical face seal , 2000 .

[8]  François Gastal,et al.  MecaNiCAL, a supply–demand model of carbon and nitrogen partitioning applied to defoliated grass: 1. Model description and analysis , 1998 .

[9]  E. Wacholder,et al.  Application of the Adjoint Sensitivity Method to the Analysis of a Supersonic Ejector , 1984 .

[10]  Lennart Ljung,et al.  System Identification: Theory for the User , 1987 .

[11]  G. van Straten,et al.  Calibration and sensitivity analysis of a dynamic model for control of nitrate in lettuce , 1999 .

[12]  Matthias Reuss,et al.  Optimal Experimental Design for Parameter Estimation in Unstructured Growth Models , 1994 .

[13]  N. Draper,et al.  Applied Regression Analysis. , 1967 .

[14]  O. Nelles.  Nonlinear System Identification , 2001 .

[15]  E. J. van Henten,et al.  Sensitivity Analysis of a Dynamic Growth Model of Lettuce , 1994 .

[16]  M. A. Scaife,et al.  The relationship between crop yield (or mean plant weight) of lettuce and plant density, length of growing period, and initial plant weight , 1976, The Journal of Agricultural Science.

[17]  Z. S. Chalabi,et al.  Sensitivity analysis of a non-steady state model of the greenhouse microclimate , 1991 .

[18]  D.G. Dudley,et al.  Dynamic system identification experiment design and data analysis , 1979, Proceedings of the IEEE.

[19]  G. van Straten,et al.  Calibration of a dynamic lettuce model for greenhouse climate control using genetic algorithms , 1999 .

[20]  Axel Munack,et al.  Design of Optimal Dynamical Experiments for Parameter Estimation , 1989, 1989 American Control Conference.

[21]  A. Huitson,et al.  Applied Statistics: Analysis of Variance and Regression , 1976 .